Foodborne hazards in food in Burkina Faso, 1990–2019: a systematic review and meta-analysis

Background: Foodborne diseases impose substantial public health burden and jeopardize socio-economic development worldwide. While accurate information on foodborne hazards is needed for informed decision in food safety interventions, such information is scarce in developing countries such as Burkina Faso. We conducted a systematic review and meta-analysis of studies reporting foodborne hazards in foods in Burkina Faso to describe the present knowledge of the situation


Introduction
Consumption of foods contaminated with foodborne hazards causes more than 200 diseases, ranging from diarrhea to cancers (World Health Organization Food safety, n.d.).In 2010, 31 foodborne hazards caused 600 million foodborne illnesses worldwide-almost one in ten people in the world-and 420,000 deaths every year, resulting in the loss of 33 million healthy life years, with the greatest per capita burden falling on the subregions in Africa.Children under the age of five years bore 40% of the burden (World Health Organization, 2015).The burden of foodborne illness is higher in the subregion of Africa with high child and adult mortality (AFR D), to which Burkina Faso belongs, compared with other subregions (Havelaar et al., 2015).
Food-producing animals (e.g., cattle, poultry, pigs and fish) are the main reservoirs for several important foodborne pathogens (Gal-Mor et al., 2014;Ferrari et al., 2019).Vegetables, water and beverages are also high risk foods for foodborne pathogens (Berger et al., 2010;Paudyal et al., 2017;Dinede et al., 2020).Moreover, produce and animal source foods consumption is expected to markedly increase in low-and middle-income countries (LMICs), raising food safety concerns unless appropriate food safety management is implemented (Grace et al., 2020).
Although consumers' concern about food safety has been increasing in LMICs, food safety management is neglected, with especially low compliance with food safety regulations in the traditional or informal markets, where most people buy food (Grace, 2015).Risk factors for unsafe food include lack of clean water used for cleaning and processing of food; poor foodproduction processes and food-handling (including inappropriate use of agricultural chemicals); inadequate food storage infrastructures; and inadequate enforcement of regulations (World Health Organization, 2015).
Several studies show that food handlers in Burkina Faso have poor knowledge about food safety (Barro et al., 2002;Ilboudo et al., 2009;Somda et al., 2018).Poor sanitary practices during food production, sale, preparation, cooking and servicing leads to microbial contamination.Studies in Burkina Faso have shown that foods obtained from retail markets are often contaminated with foodborne hazards, with food handlers in these retail markets frequently having poor hygienic practices while handling foods, including using unclean water for cleaning and processing of foods and storing food at inappropriate temperatures (Kagambèga et al., 2012a,b;Somda et al., 2018).
Food safety authorities are responsible for food safety monitoring to help safeguard consumers' health.In high-income regions, for example in the European Union (EU), responsible bodies for food safety issues set out sampling strategies and acceptable standards for microbiological hazards in foods sold, with removal from market and potential prosecution if these standards are exceeded (European Commission Regulation, 2005).Such monitoring schemes need comprehensive information on food safety, including the risk of different foodborne hazards, to inform public health policies and implement interventions that contribute to reducing foodborne disease burden.Systematic reviews reporting food hazards have been conducted in some African countries to provide synthesized information on food contamination in these countries, which help inform policy and decision making by different stakeholders (Alonso et al., 2016;Paudyal et al., 2017;Oduori and Kwoba, 2022;Gazu et al., 2023).However, such information is lacking in Burkina Faso.Given the paucity of food hazards information in this country, this Systematic Literature Review (SLR) was conducted to summarize the evidence on hazards in foods in Burkina Faso from 1990 to 2019 through systematic review of available literature reporting foodborne hazards in foods.Specific objectives of the review include describing: (1) the trends of food safety research in the review period; (2) the hazards studied including their types and associated contamination levels in the food samples and (3) to estimate a pooled contamination levels (prevalence, microbial load) of the hazards in different food samples through metaanalysis, where applicable.

Research questions
This systematic review and meta-analysis aimed to address the following research questions related to hazards in foods in Burkina Faso during the review period: a) What potential foodborne hazards (biological and chemical hazards) have been reported in foods in Burkina Faso?

Search strategy and study selection
Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline (Page et al., 2021) was used to conduct this systematic review.We searched for abstracts published in English or French in PubMed and CAB Direct databases from January 1, 1990 to September 30, 2019.We used different search terms including, but not limited to, the following keywords: "foodborne, " OR "food borne" OR "food-borne, " OR "food safety, " OR "food related, " OR "food associated, " OR "food derived, " "food* illness" OR "food* disease*" OR "food* intoxica*" OR "food* poison*, " "food pathogen" OR "food* microb*" OR "food* vir*" OR "food parasit*" OR "food toxin, " AND "Burkina Faso." Search terms kept broad and more general to allow retrieving all the literature reporting hazards in different aspects: proportion of foods with hazards, microbial loads of hazards in foods and so on.
We included studies (that followed probabilistic approach in their study design) that reported foodborne hazards in foods (prevalence or microbial load) in Burkina Faso.We excluded studies focusing exclusively on non-foodborne disease hazards and antimicrobial resistance.We also excluded studies reporting hazards exclusively in non-foods items, such as in animal feces, animal serology, or carriage in vectors.Studies outside Burkina Faso and papers on basic science were also excluded.Furthermore, studies reporting literature review findings were excluded.
Searched abstracts were imported to Mendeley to remove duplicates.Two reviewers then independently screened the titles and abstracts of the articles against the given inclusion and exclusion criteria.Articles considered relevant by both reviewers were kept and those considered relevant by just one reviewer were reviewed again by a third reviewer, to decide on their eligibility.Then, full text articles of relevant studies were retrieved.
Full text articles were assessed for their quality against the set criteria, which includes use of (1) scientifically sound methods, (2) appropriate laboratory procedures, (3) appropriate data analysis and, (4) reporting of accurate results.Articles were rated 'good' , 'medium' and 'poor' using detail criteria indicated in Table 1.Only studies rated good, and medium were considered for inclusion.
One of the two reviewers extracted the required data which included authors, publication year, title, study design, study site, sampling points, food sample types, sample size, hazard group, and specific hazards.

Data analysis
We described study characteristics using descriptive statistics (frequencies).A random-effects model (DerSimonian-Laird method) was used to estimate a pooled prevalence and microbial load, and the corresponding 95% confidence intervals (95% CI).Heterogeneity across the studies was assessed using the Higgins I 2 statistic, with 25, 50, and 75% values of I 2 showing low, moderate and high heterogeneity, respectively (Higgins et al., 2003).Sources of heterogeneity across the studies were assessed using subgroup analysis, a value of p for this test of less than 0.1 indicates a statistically significant subgroup effect (Richardson et al., 2019).Studied food samples were grouped as chicken, cattle, sheep, fish, pig, dairy products, plant, vegetable, fruit, cereal, legume, and water samples, with this classification being used for our subgroup metaanalysis, where applicable.But samples from chicken were classified as chicken meat and those from cattle and sheep were classified as meat for subgroup meta-analysis purpose.That is hazard contamination level reported in one of these subgroups was summarized as subgroup pooled estimate, which contribute to the overall pooled estimate.A study reporting a given hazard in multiple sample types was used as different records (studies) for the meta-analysis-multiple entry of a given study in the meta-analysis.Forest plot was used to visualize outputs of meta-analysis.
Prevalence estimate from a single study was kept in the metaanalysis (shown in forest plots) as it would contribute to the overall pooled estimate.The analysis was done using R (version 4.1.3,The R Foundation, Vienna, Austria), with meta package (version: 5.5-0) (Team, R.D.C, 2022).

Searching and selecting of the studies
A total of 188 unique records were identified from PubMed and CAB direct databases.Of these, 171 records were excluded based on the screening of the titles and abstracts.Full texts of the remaining 17 articles were retrieved and three articles among these were excluded during further assessment for eligibility and quality.Of these 14 articles, 36% (5/14) was rated good and 64% (9/14) rated medium using the quality assessment criteria, with 14 of them being included in the qualitative synthesis (Figure 1).

Campylobacter spp.
Of the studies included in our review, one article (Kagambèga et al., 2018) reported Campylobacter spp. in poultry feces and carcasses collected from retail markets in Ouagadougou.The study showed that 68% (70/103) of fecal samples and 50% (10/20) of carcass samples were positive for Campylobacter.

Staphylococcus
Of the reviewed studies, three studies (Barro et al., 2002;Ilboudo et al., 2009;Cissé et al., 2019) reported Staphylococcus in chicken, meat, dairy products, fruits, cereals and legumes.The overall pooled microbial load of Staphylococcus in the studied food samples was 3.2 log (95% CI: 2.8-3.6)CFU per g or ml of food, with the highest microbial load being in chicken samples: 4.5 log (95% CI: 2.8-6.2) CFU per g or ml of food (Figure 3).Our meta-analysis revealed substantial variation of microbial load of Staphylococcus across the studies (I 2 = 100%, p < 0.01).

Fungi
Waré et al. ( 2018) reported fungal spp.prevalences ranging from 3 to 70% in food samples with a pooled prevalence of 17.3% (95% CI: 1.6-73).They reported higher Aspergillus spp.prevalence (70% (95% CI: 63-76)) in infant flours than Penicillium spp.and Fusarium spp.Another study showed that 10% (95% CI: 7-14) of sorghum samples were contaminated with mycotoxins.This study also indicated that children below 6 years old in Burkina Faso have the highest mean consumption of sorghum per kg body weight in sub-Saharan African, implicating that they are a group at risk for mycotoxin exposure (Ssepuuya et al., 2018) (Figure 10).

Toxoplasma
Of the studies included in this review, one article reported a 29% (87/300) Toxoplasma gondii seropositivity in pig carcasses sampled • Aspergillus spp.

Discussion
Our systematic review finding suggests increasing activities in food safety research in the country because we found a steady rise of the studies in food safety during the review period .However, most of these studies were conducted in the capital city of the country and assessed a small number of food hazards reflecting that the research is still inadequate in the country.Our review also implies widespread contamination of foods with different foodborne hazards including bacteria (Salmonella, Campylobacter, Staphylococcus, E. coli), fungi and parasites (Toxoplasma gondii), with bacteria the most studied hazard.In Forest plot showing prevalence of Salmonella in chicken meat, meat, milk, water, beef intestine, fish, lettuce and infant flours. 10.3389/fsufs.2023.1232992 Frontiers in Sustainable Food Systems 08 frontiersin.orgaddition to the food hazards, the studies also investigated microbial loads of hygiene indicator bacteria such as Enterobacterales, aerobic mesophilic bacteria and thermotolerant coliforms in the food samples.Even if we reviewed and used the information of these hygiene indicator bacteria to support and interpret our findings on food hazards, we had not presented detail reports of these bacteria as the primary interest of this review was in foodborne hazards, but results are still reported as supplementary materials associated with this article (Supplementary materials S1-S3).The presence of these hygiene indicator bacteria in food samples suggests poor hygienic practices in food handling in Burkina Faso.Our review also indicated that foods have been contaminated with hazards along various food value chains such as beef, poultry, pork, fish, vegetables and crops, suggesting the extent of food safety problems in the country.
We found that most of the identified studies were conducted in the later decades of the review period (2010-2019), which indicates increasing efforts in food safety research.However, the sites of the studies were mostly geographically limited to the capital city of the country and only a few of the many foodborne hazards were assessed, showing that research remains inadequate.Furthermore, the studies investigated foodborne hazards in different food samples of chicken, cattle, sheep, fish, pig, dairy products, plants, vegetables and fruits, cereals and legumes, with most of the articles studying hazards in chicken meat.However, the studies often assessed bacteria, indicating limited focus on other foodborne hazards (although bacterial hazards are responsible for most of the foodborne disease burden).Consistent with our findings, food safety is often under-invested in Africa, especially in the dominant informal markets (GFSP, 2019).Food safety studies in Africa often have limited capacity-covering limited geographic areas of the study country and only a few foodborne hazards in their investigation scope (Paudyal et al., 2017).
Our present meta-analysis revealed contamination of chicken meat, meat, milk, water, beef intestine, fish, lettuce and infant flours with Salmonella, with higher pooled prevalence in chicken meats than meats and milk samples.Our review finding illustrated that meat products including-beef, mutton, chicken meat-appear to account for 44% of the pooled Salmonella prevalence reported in this review.Furthermore, subgroup analysis of our meta-analysis demonstrated that prevalence of Salmonella could be significantly affected by sample types.Salmonella have often been isolated from the gastrointestinal tract of animals, but with varying colonization levels in different hosts (Ferrari et al., 2019;Thomas et al., 2020).Chickens are asymptomatic carriers of Salmonella, leading to cross-contamination of the carcass during or after slaughtering, causing higher prevalence in poultry products (Dione et al., 2009).Poultry is the most common source of foodborne salmonellosis in the United States of America, with chicken, turkey and egg products attributing to nearly seven out of ten human cases (Thomas et al., 2020).Higher Salmonella prevalence in foods has been reported in Africa (Paudyal et al., 2017), with poultry in Africa having a higher prevalence of Salmonella than other food animals (Thomas et al., 2020).One study found that Salmonella were significantly more likely to be isolated or detected from western (including Burkina Faso) and central African food animal samples than samples from northern Africa (Thomas et al., 2020).
One paper reported Campylobacter spp. in chicken carcass, with 50% of the carcass samples being positive for the bacteria.However, relatively lower prevalence of Campylobacter in chicken meat was reported in different developing countries: 32.8% in Benin (Kouglenou et al., 2020), 38.8% in Ghana (Asuming-Bediako et al., 2022) and 26.6% in Malaysia (Sinulingga et al., 2020).Variation in Campylobacter prevalence could be due to animal species, season of the study and sample types (Ozbey and Tasdemir, 2014;Thomas et al., 2020).
Studies in this review found Staphylococcus in food products made from chicken meat, meat, dairy products, fruits, cereals and legumes, with an overall pooled bacterial load of 3.2 log (95% CI: 2.8-3.6)CFU per ml or g of food, highest load being in chicken samples: 4.5 log (95% CI: 2.8-6.2) cfu per ml or g of food.Buzon-Duran, et al.  finding of this meta-analysis demonstrated an exceedance of the maximum microbiological limit of the hazard in foods for human consumption.This systematic review demonstrated presence of E. coli in beef intestines, meat (beef, mutton), chicken meat and dairy products, with the highest prevalence in beef intestines and the lowest in dairy products.Our research found a 40% overall pooled prevalence of E. coli in the studied food products.Previous studies reported comparable E. coli overall prevalence values in foods, such as 35% in Africa (Paudyal et al., 2017) and 34% in developing countries (Mengistu and Tolera, 2020).However, other studies reported lower prevalence values, such as 15% in Ethiopia (Assefa and Bihon, 2018) and 4% in China (Paudyal et al., 2018).Differences could be due to variations in food products as, for example, in our meta-analysis we found that while meat products (beef, beef intestines, mutton, chicken meat) accounted for about 70% of E. coli pooled prevalence, the rest was contributed by dairy products.However, sub-group analysis of E. coli prevalence in the present meta-analysis indicated that sample types have weak effects on E. coli prevalence.Although E. coli is a common inhabitant of gastrointestinal tract of animals and humans and not all strains are pathogenic, some E. coli are pathogenic capable of causing illness in humans such as diarrhea or illness outside of the intestinal tract (Levine, 1987;Kaper et al., 2004).Five out of the six diarrheagenic E. coli pathotypes including STEC, EPEC, ETEC, EIEC, and EAEC were reported by the reviewed studies included in the recent review.In addition, a previous study has illustrated that an estimated burden of ETEC associated with beef, dairy, poultry meat, and vegetables was found increasing in Burkina Faso from 2010 to 2017 (Havelaar et al., 2022).The presence of E. coli in foods is an indicator of both poor hygienic handling practices of foods and of presence of other fecal pathogens in foods (Health Protection Agency, 2009).Food can become contaminated with E. coli at all stages of food production and retail.Evisceration during slaughtering and defecation during milking are critical events where E. coli are likely to enter food products destined for human consumption (Hussein, 2007).
Several factors seem to contribute to the higher contamination levels of studied food samples with the identified food hazards in Burkina Faso.Poultry is one of the main asymptomatic carriers of Campylobacter and Salmonella (Thomas et al., 2020), which might cause the higher contamination level of chicken samples with these hazards especially during slaughtering due to carcass cross-contamination.Studies in Burkina Faso have provided evidence that chicken retailers slaughter chickens in a traditional way-they themselves slaughter chickens in marketplaces, executing bleeding, plucking, evisceration, and cutting on the same table; rinse carcasses in the same bucket of water and sold off a table at ambient temperature without any type of protection from dust and pests at any point during the day.Moreover, poultry vaccinations are not mandatory in the country which leads to higher pathogen prevalence in the poultry population, causing higher carcass cross-contamination during slaughtering (Kagambèga et al., 2018).Food retailers in Burkina Faso have inadequate safe food handling knowledge and practices (Barro et al., 2002;Ilboudo et al., 2009;Kagambèga et al., 2018).Furthermore, we found, in our meta-analysis, higher microbial loads of hygiene indicator bacteria including Enterobacterales, aerobic mesophilic bacteria and thermotolerant coliforms in different food commodities representing poor hygienic practices related to food handling (Health Protection Agency, 2009), which might contribute to the higher contamination of studied food commodities with the hazards in the country.
One of the unique strengths of this systematic review is its ability to do a meta-analysis, providing pooled contamination levels of the hazards in the studied food samples.To the best of our knowledge, this is the first systematic review of foodborne hazards in Burkina Faso using a meta-analysis.A meta-analysis generates pooled estimate results through combining individual studies' results which is more informative compared to narrative reviews' findings.Such pooled estimates from a meta-analysis would enable informed policy-and decision-making (Van Wely, 2014).
However, this review was subjected to two main limitations.First, high heterogeneity was observed among some studies included in this meta-analysis, although, combining studies with low heterogeneity is recommended to be sure that the studies' findings are comparable.However, as systematic reviews synthesize results from studies that are diverse in different aspects such as sampling techniques, laboratory methods, sample size and so on, it is almost inevitable to see some heterogeneity across the studies (Higgins et al., 2002).We conducted subgroup analysis by categorizing food products that share common characteristics such as meat, chicken meat, intestines, dairy product and so on to understand the sources of heterogeneity.We found that the prevalence of some hazards such as Salmonella could be affected by sample types, reflecting that even if studies are comparable in many ways heterogeneity could arise from the sample types investigated.Strict inclusion and exclusion criteria were also used to include only relevant studies in the review.Second, results of few studies were combined in the recent metaanalysis, for example, only two studies were used in the meta-analysis of Staphylococcus.Although results from two primary studies can be combined in a meta-analysis, combining findings of more studies would improve the quality of meta-analysis findings.It is therefore recommended to cautiously interpret the findings based on the limitations.Our review found more food safety studies in recent years, indicating growing awareness of this problem.However, many important hazards receive little research attention and there is sparse information on food hazards outside the capital of Burkina Faso.In addition, most studies were at the retail point and there was a lack of information from other important nodes of the food value chains (e.g., production, processing, consumption).
Our review findings also demonstrated high prevalence of contamination of foods with hazards, including Salmonella spp., Campylobacter spp., Staphylococcus spp., toxigenic E. coli, fungi and Toxoplasma gondii.The presence of hygiene indicator bacteria such as Enterobacterales, aerobic mesophilic bacteria and thermotolerant coliforms in foods is indicative of poor hygiene practices while handling foods.These may include undercooking, cross contamination from raw food especially meat, food handlers or food contact surfaces as well as poor temperature and time control.Our findings indicated that a variety of food samples were reported contaminated with food hazards reflecting the need to target different food value chainsincluding beef, poultry, pork, vegetables, cereals, fruits and water-for food safety interventions.Consistent with our findings, a review on food hazards in Ethiopia showed that food contamination with hazards is common in beef, poultry and vegetable food value chains (Gazu et al., 2023).The widespread contamination of foods with hazards in the country raises public health concerns, especially for vulnerable groups such as children, the elderly, pregnant women and the immunocompromised.
Our results suggest that interventions are urgently needed to improve the safety of food retailed in the capital city of the country.More information is needed on food safety in outside of the capital city of the country; in different food value chains; at the different points along the food value chains and on a variety of food hazards.We suggested food safety interventions targeting different food value chains, and improved monitoring of hazards in food in the country.Forest plot showing EAEC prevalence in meat, beef intestines, chicken meat and dairy products.Forest plot showing EIEC prevalence in chicken meats and dairy products. 10.3389/fsufs.2023.1232992 Frontiers in Sustainable Food Systems 14 frontiersin.org

FIGURE 1
FIGURE 1PRISMA flowchart showing abstract search, screening, and inclusion of eligible articles reporting foodborne hazards in Burkina Faso in the systematic review and meta-analysis, 1990-2019.

FIGURE 3
FIGURE 3Forest plot showing microbial loads of Staphylococcus in chicken meat, legumes, meat, dairy products, fruits and cereals.
found a slightly higher bacterial loads of Staphylococcus in poultry meat, 4.07 ± 0.80 log10 cfu/g and Tsehayneh et al. (2021) reported comparable Staphylococcus bacterial loads in Ethiopia in raw meat from butcher shops, 3.40 ± 0.63 (log10 cfu/g).In general, the bacterial load counts requirement for Staphylococcus in food products should be below 20 cfu/g (Health Protection Agency, 2009), reflecting that

FIGURE 4
FIGURE 4Forest plot showing E. coli prevalence in beef intestines, meats (beef, mutton), chicken meats and dairy products.

FIGURE 5
FIGURE 5Forest plot showing STEC prevalence in meats, beef intestines, chicken meats and dairy products.

FIGURE 6
FIGURE 6Forest plot showing ETEC prevalence in meats, beef intestines, chicken meats and dairy products.

FIGURE 7
FIGURE 7Forest plot showing EPEC prevalence in meats, beef intestines, chicken meats and dairy products.

TABLE 1 Quality assessment checklist for articles included in the systematic review and meta-analysis of food hazards in Burkina Faso, 1990-2019.
analysis Appropriate statistical methods were used for data analysis Acceptable statistical analysis but with some limitations.Limitations were acknowledged and accounted for.Inappropriate statistical analysis 4 Accurate results Results were complete and accurate Results remain valid Results were incomplete or inaccurate Dinede et al. 10.3389/fsufs.2023.1232992Frontiers in Sustainable Food Systems 04 frontiersin.org Reviewed studies in the systematic review and meta-analysis of food hazards in BurkinaFaso (1990Faso ( -2019) )with their study sites, sampling points, sample types and food hazards investigated.